00:00:00 --- log: started forth/04.07.04 00:39:03 hey, I wrote some documentation! 00:39:40 the cool thing about writing documentation is you don't have to debug it so much 00:46:17 --- join: qFox (C00K13S@cp12172-a.roose1.nb.home.nl) joined #forth 00:54:23 --- join: crc (crc@0-1pool176-6.nas6.philadelphia1.pa.us.da.qwest.net) joined #forth 01:31:17 --- join: MorgothDc (~morgoth@gw-bdi-3.ticom.pl) joined #forth 01:34:32 * crc greets MorgothDc 01:35:26 Hell[o] 01:38:14 --- join: Serg (~z@193.201.231.126) joined #forth 01:41:06 * crc greets SDO 01:41:13 s/SDO/Serg/ 01:41:15 :-) 01:41:50 hi 01:42:34 yesterday i had a great battle w/ poorly designed website ^) 01:44:02 * crc hates poorly designed sites 01:44:18 me too, but i won finally 01:44:45 What was your battle? 01:44:48 i cleared the URL's and uploaded PhpMyAdmin to hack SQL database 01:45:27 * crc hates SQL databases 01:45:38 the damn thebmaster was demanding $100 for writing "add publication" script 01:45:50 now they can do it for free ;) 01:45:55 Good 01:46:03 Free things are good 01:46:26 * crc is doing some web programming with RetroForth :-) 01:46:28 this is the major lawer company engaged in some "loud" lawsuits ;)))) 01:46:45 I see 01:47:04 I'm not fond of lawsuits, except in extreme cases 01:47:17 ame too 01:48:29 I'm trying to figure out how to use RetroForth from within Apache :-) 01:48:48 i came to them for free consult (graduates offer this, as learning practice), helped a little w/ Win98 ... 01:49:28 good lawers are valuable people to make friends with ;)))) 01:50:01 --- quit: MorgothDc (Read error: 104 (Connection reset by peer)) 01:50:30 why not wider useg gforth ? 01:50:33 mod_forth, he-he ;) 01:51:00 1) I don't like gForth 01:51:27 2) I can't recompile gForth on the remote server 01:52:02 3) I can't add mod_forth as I don't manage the server 01:52:15 so how do u script in Forth ? 01:52:25 #!/usr/bin/myforth 01:52:26 ? 01:52:46 Right now I have a PHP script that calls RetroForth, redirecting the page source into the interpreter 01:52:46 no, no 01:52:58 RetroForth doesn't support command line arguments 01:53:27 i say write mod_forth on your home machine for general use ;) 01:53:44 or does it exist already ? 01:54:17 I'm not doing this for my home computer! I want to use RetroForth to generate the HTML for http://www.retroforth.org 01:54:35 hmm... 01:55:06 does it have powerful string manipulation library, what is the must for webscripts ? 01:55:26 No powerful string manipulations in RetroForth 01:55:46 hmm.. so how do u script w/o so ? 01:56:54 http://www.forthfreak.net/wiki/index.cgi?WikiWrittenInForth 01:57:22 won't follow - Xpensive traffic ... 01:57:46 --- nick: Serg -> Serg[GPRS] 01:57:56 I have a PHP wrapper that calls RF and redirects the source page into it 01:58:18 ok, ok 01:58:20 system("./rf <".$page); IIRC 01:59:16 --- quit: Herkamire ("lots of progress on herkforth tonight. now time for bed") 01:59:19 * Serg[GPRS] can't imagine script w/o regex, split and Perl hashes 02:00:56 --- join: crc_ (crc@0-1pool176-4.nas6.philadelphia1.pa.us.da.qwest.net) joined #forth 02:01:10 my scripts are pretty simple 02:01:17 --- quit: crc (Read error: 104 (Connection reset by peer)) 02:01:22 --- nick: crc_ -> crc 02:02:15 I'm basically using a simplified, wiki-like, markup language written in RetroForth and want to convert it to HTML on the fly 02:02:29 aha 02:02:38 * crc wants Apache to call RetroForth, redirecting the page name as input for .FML pages 02:02:57 i'd rather use XML 02:03:30 XML is a bloated piece of crap 02:03:31 * Serg[GPRS] likes to invent funny tags 02:03:56 Since the markup are written in Forth, I can add new ones at will anyway 02:05:00 * Serg[GPRS] eating, pause 02:11:11 It's mostly working now :-) http://www.retroforth.org/RF.fml 02:35:01 --- quit: Serg[GPRS] (Read error: 110 (Connection timed out)) 03:11:40 The FML code has now been updated to work with the Apache loaded scripts :-) 03:17:35 ." Hello from " * ." RetroForth 7.0 " * ." and " / ." FML!" / cr 03:17:50 * crc just pasted into the wrong window :-( 03:59:19 --- quit: aum_ () 03:59:25 --- quit: arke ("VACATION BYE 3 WEEKS!!") 04:57:18 --- quit: crc ("Sleeping...") 05:43:31 --- nick: lalalim_ -> lalalim 05:49:50 --- join: tathi (~josh@pcp02123722pcs.milfrd01.pa.comcast.net) joined #forth 06:30:38 --- quit: SDO (Read error: 110 (Connection timed out)) 07:47:48 someone on linux wanna help me out plz? 07:47:57 need something built 07:47:58 With? 07:48:14 * Robert hides. 07:48:25 oh no, i saw that! 07:48:51 according to this guy, it should be piss easy 07:51:41 :) 07:51:50 why don't you do it yourself then? 07:51:55 i dont have linux 07:52:08 appearantly this stuff needs to be assembled on linux 07:52:12 i dont know :\ 07:52:25 so...then what are you going to do with it? 07:52:26 i just want a bootable ciforth >:( 07:52:37 the thing ought to create a bootable image 07:52:42 ah, ok. 07:52:51 linuxfree! :p 07:58:48 well, if I had an x86 box I'd help :) 08:00:03 Hrm. 08:00:09 Well, I'll make it if it's easy 08:00:19 you dont have a x86 box? 08:00:20 :\ 08:00:25 how's that possible? :) 08:00:30 Robert> k, i'll send the package 08:02:57 qFox: I have a dual-G4 PowerMac running Gentoo GNU/Linux. That's my only machine. 08:03:03 I'm quite happy to have it that way :) 08:03:44 * Robert looks around 08:03:56 10 machines. But nothing that sexy. 08:04:10 otoh, nothing that expensive either ;) 08:05:09 i have um, an intel, a celeron, and a bunch of 486 and p33/p66's 08:05:15 i believe all x86 :p 08:05:35 unless my laptop isnt, its compaq, you never know what they might have done with it :p 08:09:36 --- join: fridge_ (~fridge@dsl-203-33-165-169.NSW.netspace.net.au) joined #forth 08:15:17 I got a few non-x86 computers. 08:15:33 And one x86-but-non-IBM-compatible one, using the i80186. 08:25:09 --- quit: fridge (Read error: 110 (Connection timed out)) 09:57:28 --- join: kc5tja (~kc5tja@66-74-218-202.san.rr.com) joined #forth 09:57:33 --- mode: ChanServ set +o kc5tja 09:58:02 Hi 09:58:06 Howdy. 09:58:20 Just doing some work on the Kestrel. 09:58:35 I'm thinking it's time to switch to blocks for source code now that the emulator is halfway stable. 10:00:10 What did you use before? 10:00:24 The emulator and its assembler are written using files. 10:00:39 And for now, will remain as files. 10:01:13 But since the code governing the text cursor will inevitably make it into the Kestrel Forth ROM image, I think it's best to store it in a format that Kestrel Forth itself can read. :) 10:05:01 OTOH, there is a certain convenience that files do afford: the ability to keep the OS sources in a source code management system. 10:07:14 Actually, you know what? 10:07:20 I think I *will* eschew blocks in favor of "files." 10:07:53 In the same spirit as blocks, I'll treat the disk space as a virtual address space (same basic goal as orthogonal persistence). 10:08:18 But instead of rigid 1K blocks in that address space, I'll treat it as a flat address space. 10:08:43 To store data in the space, you need to allocate a chunk from that space (chunks are always contiguous, therefore, as with the result of "malloc" in C). 10:10:04 Individual objects in the disk image would be referenced by a unique integer -- "formatting" the disk involves just clearing the disk's object table. 10:10:25 So an entire application's code can be loaded with something like 5 LOAD. 10:10:55 where 5 refers to object #5, not block #5. 10:12:14 To load an object into memory for manipulation, you'd say something like "5 LOCK", which locks object 5 into RAM. 10:12:49 ciforth uses a similar system, except everything is loaded from one "file" (in the end the result is the same i guess) 10:13:03 You manipulate the object, then call "DIRTY" to mark the object as having been modified. 10:13:42 Then "5 UNLOCK" would unlock the region from memory. If a swap-out needs to be done, the status of the dirty bit would be used to determine whether a write-back is necessary. 10:13:57 --- join: Herkamire (~jason@h000094d30ba2.ne.client2.attbi.com) joined #forth 10:14:05 qFox: let that one file be the whole harddrive. 10:14:11 yeh 10:14:18 And ciForth isn't unique -- PC/GEOS was doing this back in 1983. 10:14:26 And it is a **FAST** OS. 10:14:39 I'd say GEOS on a 286 compares easily with a Pentium II box running X11. 10:16:49 --- join: fridge (~fridge@dsl-203-33-162-127.NSW.netspace.net.au) joined #forth 10:17:29 And yes, GEOS stored its "virtual memory" images in files. In fact, "saving" a file in GEOS simply flushed the VM state to the VM file. 10:17:34 oh no i didnt mean it was the first, i'm just saying.. 10:17:43 "Opening" a file simply changed VM files. 10:17:45 Yeah, I know. 10:17:57 I'm just saying -- GEOS had a damn advanced persistence model. 10:19:07 :) 10:21:33 persistence is cool 10:22:00 Herkamire: I'm thinking of changing to an object database-style of storage for my Forth instead of using plain blocks. 10:22:12 kc5tja: :) 10:22:34 cool 10:22:36 This solves several nagging problems. 10:22:53 First, it's more flexible, but at the added cost of code complexity in the Forth kernel. 10:22:53 which problems? 10:23:06 --- quit: fridge_ (Read error: 113 (No route to host)) 10:23:16 If I were to write a web browser in Kestrel Forth, for example, I need some way of storing web pages. 10:23:30 I'm having some issues with blocks, and I'm curious if others have he same 10:23:33 Problem is, I'll also need some way of storing images, etc. 10:24:20 All in all, even if web page sizes are all given in advance, it's still a very difficult problem to determined _a priori_ how much space to allocate for storing the page data during viewing. 10:25:29 Therefore, what's needed is a storage system that can accomodate variable length, uncertain, potentially even *growing* amounts of storage on-the-fly. 10:25:38 yeah 10:25:57 it seems to me something like that would make some programming project monumentally easier 10:26:24 I'm also looking to store source code in this format as well. 10:27:04 right 10:27:05 And think of the benefits too: the ability to directly reference a *raw* binary image (e.g., font data or icon data) right in the source code itself (something I've *ONLY* seen implemented for GEOS in the Commodore 64!!). 10:27:32 :) 10:27:33 : QuitIcon _actual picture here_ ; 10:27:49 yeah :) :) 10:27:54 herkforth will be like that more or less 10:28:49 I don't know if I will show a definition for data anywhere, but when you put your cursor on the name of the date (ie anywhere it says QuitIcon) it will display the data 10:29:50 I hope to have a zooming interface before I do pictures, so that's a ways off, but it's not that far off to have text and tables working this way 10:30:26 I actually am thinking that I don't want a full-bore ZUI. 10:30:43 I want a gigantic, variable-sized desktop -- e.g., a desktop that effectively spans an infinite plane. 10:30:58 All content on the desktop appears at the same depth. 10:31:20 But to navigate, I can scroll the mouse past the edges of the screen, OR, I can zoom out to get a bird's eye view. 10:31:47 why don't you want to do ZUI? 10:31:53 It's too hard. 10:32:24 ahh 10:33:00 It also requires floating point arithmetic (or, at the very least, costly double-precision integer arithmetic) to get any kind of precision, and even THAT is finite. 10:33:16 right 10:33:23 I haven't thought about that yet 10:33:37 So my desired desktop is more a compromise solution. 10:33:46 Panes, panels, tiles, and lots of physical space. :) 10:34:05 cool 10:34:41 ion is pretty cool. I really don't have much to complain about 10:36:09 The only think I complain about is it shanghai's too many keyboard controls. I always have to switch out of Ion and into Amiwm whenever I want to run any applications that depends on lots of control key combinations that Ion subsumes for itself. 10:36:39 Remapping the keys in Ion only partially solves the problem. It's not a total solution. 10:36:51 control keys? 10:37:10 all my ion keys use alt.... I probably remaped all/most of them though 10:37:24 Control keys in the controlling sense of the term, not CTRL. 10:37:56 the only time I have trouble with ion and software vying for keys, is when software wants to use F-keys 10:37:57 If I could somehow get Ion to use the Windows keys instead of ALT, that would be of extremely great help. 10:38:16 yeah, I want ion to use the apple key :) 10:38:43 I think I did that with sawfish 10:42:02 Hmm...I can't possibly think of any additional words for this object system to use. 10:42:23 * Tomasu is back (gone 12:06:41) 10:42:25 NEWREC ( u -- n ) -- Allocates a new record of size 'u', and assigns to object ID 'n'. 10:42:33 DISCARD ( n -- ) -- Discards record 'n'. 10:42:49 LOCK ( n -- a u ) -- Locks record 'n' in RAM; returns its current address and size. 10:42:59 UNLOCK ( n -- ) -- Unloads record 'n'. 10:43:16 DIRTY ( n -- ) -- Marks record 'n' as being modified. 'n' must be locked; otherwise, this is ignored. 10:43:48 Maybe a GROW and SHRINK word to change the size of a record. But that's *it*. 10:45:14 cool 10:46:42 Of course, should an error occur (e.g., attempt to grow a record too large, or not enough RAM to lock a record), how to handle that? 10:47:23 right 10:47:33 I'll probably just use call-backs to implement handlers for those cases. 11:02:05 Does libc preserve any (which?) registers on x86? 11:02:55 I think it does, but I don't know which ones. Your best bet is to assume all registers, except for EAX and EDX, are modified by libc. 11:03:13 EDX:EAX is of course modified, but it's also the C return registers. :D 11:03:43 OK, thanks. :) 11:09:55 Damn, time to get some food. 11:26:15 --- join: SolarFire-| (SolarFire@pD9545D0F.dip.t-dialin.net) joined #forth 11:43:42 back 11:44:01 Hi again 11:44:03 --- quit: SolarFire (Read error: 110 (Connection timed out)) 12:05:33 Actually, I kind of thought of a different idea. 12:06:34 Let's assume we have normal "blocks". For the sake of argument, let's further say a block is 1024 bytes long, as it currently is. 12:06:53 Prepare for flashbacks to the Commodore 64 running GEOS, because this is essentially the VLIR file concept. 12:07:58 We can identify a file as being a block which refers to other blocks. A single 1K block, using 4-byte pointers, can reference 256 other blocks containing data -- thus, a one-deep hierarchy would allow files up to 262,144 bytes. 12:08:39 Another way of looking at this, though, is that a file consists of 256 1K records. 12:08:58 If we allow these records to be variable length, then we end up with a case where a single file is allowed to have 256 records of effectively infinite length. 12:09:24 To make things very, very simple, we load the entire record into memory, and if necessary, write it back out in its entirety. 12:12:04 Thus, we can create fairly sophisticated structures of arbitrary size by creating a sequential chain of blocks for each record. 12:12:28 There still would need to be a way to monitor which blocks are actually in use or not though 12:13:03 (don't mind me -- I'm just thinking "out loud.") 12:13:12 kc5tja: sounds good 12:13:30 hey folks 12:13:39 you've gotta implement a filesystem on top of blocks one way or another 12:13:49 (as that's how disks work) 12:13:55 Yes. 12:14:35 But, my goal trivially simple to implement as possible, while still retaining as much usability as possible. 12:17:18 Damn, I can't type anymore. :( 12:18:11 http://www.zimmers.net/geos/docs/vlirfile.txt -- at any rate, this is a page describing the C64 GEOS VLIR file structure. 12:29:50 --- join: crc (crc@0-1pool88-109.nas48.philadelphia1.pa.us.da.qwest.net) joined #forth 12:31:23 --- join: blockhead (default@dialin-738-tnt.nyc.bestweb.net) joined #forth 12:32:41 re crc, blockhead 12:33:04 * blockhead performs a crc :D 12:33:10 hey kc5tja 12:34:14 hello kc5tja 12:34:20 crc: you hack retroforth right 12:34:39 how so? 12:35:39 * crc hopes he's getting finally it right 12:35:57 it was a question :) 12:36:20 Yes, I hack RetroForth 12:36:35 (Put a ? after the questions) 12:39:20 * crc only uses/hacks RetroForth, except when answering questions about other Forths 12:39:27 --- join: Sonarman (~matt@adsl-64-169-95-20.dsl.snfc21.pacbell.net) joined #forth 12:39:54 Hello Sonarman 12:40:02 hi crc 12:40:13 and everyone else :) 12:40:58 --- join: crc_ (crc@0-1pool176-33.nas6.philadelphia1.pa.us.da.qwest.net) joined #forth 12:41:07 * crc_ hates dialup 12:41:24 understandable 12:42:41 --- quit: crc (Nick collision from services.) 12:43:02 --- nick: crc_ -> crc 12:46:52 Is there a faster (free) x86 emulator than Bochs? 12:48:43 i don't know how complete it is, but there's plex86, which is a virtualization engine like VMWare: http://plex86.sf.net/ 12:51:49 --- join: proteusguy (~proteusgu@69.79.24.31) joined #forth 12:52:07 I'll try it 12:59:20 Only supports Linux :-( 12:59:45 What system are you on? 13:00:03 Windows right now 13:00:10 You might want to try DosBox too -- I'm not too sure how good it is though. 13:00:17 Ok 13:00:28 Well, let me clarify. 13:00:42 It appears good enough to run PygmyForth and, believe it or not, GEM. :D 13:00:45 And it runs Windows 3.1. 13:00:49 Wow 13:00:57 But as to how fast it is, that I can't judge. 13:01:01 GEM doesn't get emulated well in my experience 13:01:10 its too slow for games though 13:01:14 Though it was a nice system back in the day... 13:01:20 crc: GEM emulates just fine. 13:01:25 The problem is intrinsic in GEM itself. 13:01:38 GEM actually *BUSY WAITS* to see if you single- or double-click an icon. 13:01:41 Which is retarded. 13:01:45 That's the *one* thing I hate about GEM. 13:02:15 * crc spent several days getting GEM/3 running under FreeDOS a couple years ago 13:02:30 GEM ran under DosBox out of the box, no problems. 13:02:40 I have it on my box. :) 13:02:52 Although, I don't have any really good reason to use it -- no apps for GEM. 13:04:47 Oooh, I am *such* a cheater. :D 13:04:51 * crc is planning to set up copies of Syllable, ReactOS, and other OSes to test RetroForth under 13:04:53 Why? 13:05:28 ': cells if, b@, !a+, 1 #, -, cells ;' then, ;' 13:05:39 Because I use the CPU's B register as a local variable of sorts. 13:10:56 :-) 13:11:29 * crc has been writing a replacement for SEDForth inspired by your target compiler 13:13:18 damnit. this denotation stuff is embedded into the core of all of ciforth. no way in hell i'll get it out unless i want to do a complete rewrite. gah 13:13:58 * slava got GLcolorforth running. 13:14:04 qFox: What actually *is* that whole denotation thing? 13:14:13 ah well i've finally figured that out 13:14:38 the author uses denotations to do stuff like accepting negative numbers 13:14:39 Well, note to self: explicit "loops" are twice as fast as counted loops in FTS1001 machine language. 13:14:39 -500 13:14:43 +3543 13:14:56 also an alias for S" is "string" 13:15:03 OK 13:15:05 attached words, so to speak 13:15:17 my problem is that it completely borks up search results 13:15:41 if the word you're looking for is not found, it'll go search in the denotation vocabulary for "a shorter match" 13:15:54 so if you make a typo, it might still reply something 13:16:01 very annoying 13:16:12 So, " is a word that maps to S"? 13:16:23 no i dont believe it does, hold on 13:16:37 heh, in fact S" maps to " 13:16:57 : S" POSTPONE " ; 13:17:22 ' is in this voc as well, so are ^ and & 13:17:38 OK, that's all very confusing. 13:17:39 its like a big weed 13:17:43 yeah well 13:17:57 at least i understand now, and will just try to live with it 13:18:13 its not terribly bad, although its not very forthish i think 13:18:33 "whatever" doesnt feel natural to forth 13:18:35 imo 13:18:40 * kc5tja nods 13:18:57 but its not really an obstruction i guess 13:19:12 just dont like 0-9 and A-F being in the dictionary 13:19:52 but tearing it out will ruin NUMBER or (NUMBER) or whatever 13:20:08 So, if you have a number like "156", and of course 156 isn't found, then it will execute word "1", which parses the "56" out and understands to push the value 156 on the stack? 13:20:11 kind of funny 13:20:31 i dont know exactly, but yeah it will execute 1 13:20:46 and possibly parse the rest and execute 5 and 6 as well 13:20:58 It won't make any sense to execute 5 and 6. 13:21:01 : 1 13:21:01 -1 IN +! (NUMBER) POSTPONE SDLITERAL 13:21:01 ; IMMEDIATE ( DENOTATION) 13:21:11 Oh, I see. 13:21:29 So it backs the input pointer to the '1' again, and just executes (NUMBER). 13:21:43 yeah 13:22:06 which makes me believe that it cannot handle BASE>17 13:22:21 Probably not in any meaningful way. 13:22:21 (ok its not required, but...) 13:22:48 s/meaningful/convenient/ 13:22:50 so the main loop, probably doesnt even contain number, just parsing and executing. 13:23:21 Almost certainly true. 13:23:37 It sounds like he tried to over-optimize his interpreter loop. 13:23:48 possibly 13:24:01 * fridge bangs his head on the desk 13:25:06 S[ ] OK 256 BASE ! 13:25:06 S[ ] OK YU 13:25:06 YU? ciforth ERROR # 12 : NOT RECOGNIZED 13:25:56 i dont get it though. why would you really want to optimize your outer compiler 13:26:03 I have no idea. 13:26:06 the only time you use it is for defining words 13:26:14 its not like it matters for running programs 13:26:15 It seems to me that he was just toying with an idea. 13:26:25 he seems to be quite successfull tbh 13:26:44 "string" does LOOK nicer then S" string" 13:27:17 i remember somebody in here was talking about a similar idea a while back. dont remember who though 13:27:29 Sure as hell isn't me. :) 13:27:33 hehe 13:27:40 ohwell. 13:27:59 * qFox crosses "remove denotation from ciforth" from todo list :p 13:28:55 how much space did 32bit cover? addressing wise 13:29:39 In terms of 8-bit bytes, 32-bit addresses can cover over 4.2 billion of them. 13:29:44 i'm wondering whether this LOAD word will load 1k block x from the drive it was booted from, or that it can do only a limited number of blocks 13:30:33 S[ ] OK SEE 468B 13:30:33 : 4 13:30:33 -1 IN +! (NUMBER) POSTPONE SDLITERAL 13:30:33 ; IMMEDIATE ( DENOTATION) 13:30:42 this is waht i mean. 13:30:48 crc: Sorry, I knew you reported about the inspiration of my cross compiler, but I was engrossed in the other conversation. If you need any questions answered about the cross compiler, I'll be happy to help. 13:31:38 qFox: This reminds me of the "catch-all" e-mail accounts that we're *constantly* trying to get our customers at work to NOT use. Then they wonder why their mail accounts get loaded with 100MB+ of spam on an almost daily basis. 13:31:49 hehe 13:31:57 (for some of them, this is no joke; they routinely bring their segment of the network down. :() 13:32:09 what could the B suffix mean, its probably used for NUMBER 13:32:27 I just see it as a hexadecimal number. 13:32:38 but the b? 13:32:44 $468B 13:32:49 ohhh 13:32:52 d'oh 13:33:12 i forget that ciforth always shows SEE in hex 13:33:12 But if the base is 10, I doubt that (NUMBER) would have anything to do with that "B". :D 13:33:30 nah SEE always shows in hex, no matter what BASE is 13:37:34 I should setup a whitelist filter for my email 13:38:12 pity I can't for work 13:38:45 I'd like to use a Bayesian mail filter that works on word pairs instead of individual words. 13:40:26 Well, I have a cursor which is fixed at location (0,0) on the screen in the Kestrel. Hardly very impressive. 13:40:55 does it accept input? 13:40:55 :) 13:41:32 No. 13:41:59 the emulator has a facility for a really dumb, super fake UART which I use for testing and debugging purposes, but eventually it'll take its input from the SDL window itself. 13:42:46 It doesn't even do interrupts yet. 13:44:44 --- quit: tathi ("leaving") 13:45:40 hm, what will i like better. an OK that shows the entire stack, or a . for every stack item 13:47:23 hm you also have to get used reading his sources: 13:47:24 : DO-DEBU 'NEW-OK >DFA @ 'OK >DFA ! 'NEW-THRU >DFA @ 'THRU >DFA ! ; 13:47:52 appearantly, this works :p 13:48:02 DEBU? 13:48:13 debug 13:48:21 my bad 13:48:47 changes the OK to show all stack items (just does S. before printing OK) 13:49:42 --- nick: Tomasu -> Typosu 13:50:28 --- nick: Typosu -> Tomasu 13:53:13 shall i use this denotation style of programming, or keep away from it... 13:53:51 Personally, I would prefer to stay away from it. But that's just me. 13:54:58 hm its tempting. 13:55:04 its there, i cant get rid of it... 13:55:51 at least i know understand what 'INTERPRET meant (i saw it in a SEE once, but i couldnt figure out why i couldnt find it :) 13:58:51 'nother reason why this is so confusing btw 14:01:39 that code is very silly 14:01:47 why not just add ".s" to the word that prits "OK" 14:02:38 Herkamire: I see a pattern with his coding style: *heavy* use of vectored execution. 14:02:53 It looks like he's striving to eliminate any and all occurances of the word "IF" from his source code. 14:03:07 who's coding is this BTW? 14:03:16 The author of ciforth. 14:03:19 i can concur with that. 14:03:22 its... 14:03:23 Whoever that is. 14:03:32 albert van der horst 14:03:36 sounds aweful 14:03:39 dutch d00d 14:03:52 sounds like it makes it very very hard to know what code exactly is being executed 14:04:09 hm, now i figured htis denotation stuff out, its ok 14:04:25 just not truely forth 14:04:46 vectors can be useful, as they provide a layer of abstraction, but I don't like it when they hide that layer 14:05:19 i'm not sure why he's changing THRU for this though 14:05:43 oh wait. i do. its for LOAD 14:05:45 nvm 14:05:51 Maybe he's printing out the current block being loaded. 14:06:03 in debug mode, the header of the block is printed 14:06:08 header being the first line 14:06:23 Yep. I have a word INDEX that prints the first lines of a range of blocks. 14:06:35 but i'm circumventing that block file and merely using INCLUDED in the windows port of ciforth 14:06:44 I use it as one would use DIR in DOS; it works amazingly well. 14:07:11 kc5tja: cool :) 14:16:40 OK, here's the deal. 14:17:11 I'm going to provide a binary-only image of Kestrel Forth, basically called "OS System 0". It's sole purpose is to bootstrap the first real version of the Kestrel's first OS. 14:18:32 kc5tja: how's the text cursor coming along :-)? 14:18:45 It's not. 14:19:32 It WOULD be coming along if I didn't have to god damned mother fucking debug variable access code every fucking time I want to use one. >>:( 14:19:55 --- join: Topaz (~top@spc1-horn1-6-0-cust217.cosh.broadband.ntl.com) joined #forth 14:21:22 here is the new factor parser, written in factor: 14:21:22 : parse ( str -- ) 14:21:22 [ end? not ] [ scan word, ] while ; 14:22:36 slava: you and your fancy lists... 14:23:15 : parse scan word, end? if; parse ; 14:23:58 it simplifies the interpreter 14:24:07 ("if;" is equivelent to "if exit then") 14:25:04 ??? your definition is almost twice as long as mine. how is that simpler? 14:25:52 your definition has 9 words/symbols in it. mine 5 14:27:25 --- quit: crc ("http://www.retroforth.org") 14:27:36 but your definition of the control structure words has to do assembler etc 14:31:22 huh? 14:31:59 WOOOHOOO!!!! 14:32:05 Herkamire: by the way [ end? not ] counts as 1 'symbol' 14:32:07 I got a keyboard controllable text cursor. 14:32:12 kc5tja: excellent! 14:32:17 slava: no it doesn't. that's 4 14:32:25 kc5tja: yay! 14:32:36 --- quit: blockhead (Read error: 113 (No route to host)) 14:32:43 ok, so mine requires assembler capabilities. 14:32:54 I don't understand how you can have a forth without this 14:33:03 Herkamire: well in memory that definition of 'parse' is a list with 3 elements, first 2 elements are lists, last one is 'while' 14:33:24 Herkamire: i have a small kernel written in C (well 'small' being 1,400 lines of C) 14:34:13 --- join: crc (crc@0-1pool176-1.nas6.philadelphia1.pa.us.da.qwest.net) joined #forth 14:34:15 It's amazingly quick too. It takes about four seconds or so to clear the screen on my box (to be expected when you have a 640x480, 256-color display for a CPU that is only running about 0.5MHz), but actual cursor updates on the screen are amazingly rapid. It is about as fast as a Commodore 128 in 80-column mode. 14:34:18 right, and the first two list elements in turn are a list of two elements. so you have 3 nodes, and 5 leaves. 14:34:38 Herkamire: the kernel provides linked lists, arrays, words, numbers. factor code cannot directly access memory it can only use the objects provided by the kernel. 14:35:07